THE GYMNOTID EELS OF TROPICAL AMERICA. 181 



scars are clear and well defined, and the new tissue quite distinct from the old, 

 conditions that would probably not obtain were the injuries old ones. 



The small number of individuals of this species showing any mutilation could 

 be explained in either of two ways. First, it is possible that injuries of greater 

 severity than the slight mutilation of the caudal appendage were fatal, and conse- 

 quently no specimens collected showed these; or secondly, the species G. carapo 

 is not frequently injured. The first explanation evidently does not obtain, for, two 

 specimens previously mentioned (see Figs. 19 and 20) had each not only lost the 

 entire caudal appendage, but a considerable portion of the body as well, and 

 more than merely surviving these injuries, had restored by regeneration much of 

 the part lost. Concluding that G. carapo is only occasionally injured, the causes 

 for this immunity are to be sought, and the color-pattern and the predaceous habit 

 of the species present themselves as probable reasons for their exemption from 

 injury. 



In life this fish is strikingly marked with numerous bands of greenish blue, 

 which cross the body at right angles. An isolated individual is thus rather con- 

 spicuous, yet these same bands may afford a certain protection when this species 

 lurks among the vertical stems of the calladium, the "Mucka Mucka" and other 

 aquatic plants, forming its normal habitat in the small streams which it frequents. 

 It was observed to be a very rapid, vigorous swimmer, and the contents of the 

 stomachs examined, as well as the shape of the head and jaws, show this fish to be 

 predaceous, at least as an adult. Both factors probably contribute to the immunity 

 of this species. Plate XX, fig. 4 is an outline drawing in which the seven injuries 

 discussed in connection with this species are indicated by outlines of the injuries. 



Sternopyginae. — The species of the SternopygincE are distinctly compressed 

 and more or less elongate. The caudal appendage is slender and well developed, 

 being almost a third as long as the entire body in some species. Two of the genera, 

 Eigenmannia and Sternopygus, have patches of very small teeth in each jaw, while 

 the others are toothless. All the species of the subfamily feed largely upon plank- 

 ton, insects, and worms. One of the first records of regeneration in this subfamily 

 is by Kaup, 1856, who prefaces his discussion of the genus Rhamphichthys with the 

 following: "One perceives sometimes at the point of the damaged tail a projecting 

 cuticular process destitute of vertebrae which resembles the reproduced tail of a 

 lizard. Judging from the narrowness of the toothless mouth these fish must sub- 

 sist on small insects and be themselves destroyed or injured by predaceous fish 

 whence it happens that the point of the tail is often defective." 



From the following table it is plain that regeneration is common among 

 these species. The injuries were of two sorts: (1) part of the caudal appen- 

 dage sometimes with more or less of the caudal region of the body had been 

 removed; (2) V-shaped, or semicircular, pieces of the anal fin and its muscles had 

 been taken out of the ventral region, often well ccphalad. The particular cases 

 are considered in the second part. The regeneration in many cases was quite com- 

 plete, muscle, fin-rays, scales, pigment, and caudal appendage having been restored. 



